Instructions and criteria for A-level project

↵ Back to class homepage

For this project, you must carry out a small research project that involves collecting data using at least one of the linguistic study techniques discussed in this class. This will most likely involve phonetic analysis, corpus analysis, or collection of reaction time data using experiment control software; the other techniques (eye-tracking and electrophysiology) are not feasible to do in a single semester.

Your project may be related to your own PhD dissertation project, or it may be something different. It may also be a replication of an existing study.

A report of a research project needs to motivate the research question and explain why it's important to know the results of this study. (This is the case even for a direct replication project: the report still needs to explain why the question tested in the original study is important, and why the study should be replicated.) Therefore, you will probably need to read and describe some previous literature. You do not need to summarize every paper ever written on this topic, but you should at least briefly describe what is already known about this question, where there is a knowledge gap (i.e., a question that previous research has not yet answered), and how your study will address that gap.

You should also explain why this question is important. For example, sometimes it can be acceptable to do an experiment that is very similar to previous experiments but in a new language (e.g., you might do a self-paced reading experiment that is directly based on one that was done with English-French bilinguals, whereas you will do it on Cantonese-Mandarin bilinguals, and then the "new" aspect of your experiment is that you're testing it on a pair of languages that hasn't been studied before using that method), but you have to explain why it's important to test that pair of languages (you can't just say that you want to do the experiment on another language without explaining why; there should be some specific interesting reason why you think the results could be different or similar in this language, and why that will teach us something important). When explaining what has been shown in previous experiments, do not plagiarize others' work.

After raising a research question and describing how you would test this in a study; you should then describe all the details of how you carried out the research (e.g., how many volunteers will participate in the experiment, how many words you used, how you analyzed the data, etc.). The methods section must be specific: if somebody reads the report, they should have all the information they need to repeat the study in exactly the same way you did it.

Criteria for earning credit

To get credit for completing the proposal, your report must meet the following criteria.

• Your report must clearly address the following questions:
  • What is your research question? This should be specific; ideally, it should be expressed as a yes-or-no question, or two possible options. Saying something like "The study examined how bilinguals process words" is not a clear research question, because it's not clear what the possible answers would be. A good research question would be something like "The study examined whether bilinguals recognize words in their first language faster than words in their second language" (because this research question implies the possible answers: either they do, or they don't), or something like "The study examined how bilinguals recognize words in their second language; one possibility was ...., and another possibility was...").
  • Motivate the research question. Why is it important to address this question? What new thing will it teach us about language or about the mind? Why is previous research insufficient for answering this question (what is missing from the previous research)? Typically the best way to report research is to start from a big-picture issue or question (like "how do people learn a second language", which is too big a question for us to address in one paper, and is not a specific research question per the above criteria) and then slowly describe how you work that down into something more specific and manageable (e.g., "How do people learn a second language.... one theory about this suggests that X.... one way people have tested that is by comparing Y and Z..... but those studies haven't considered M, which might be important because of N.... my study will examine that."). The important thing is that there must be a clear line of logic from the big-picture question down to the specific question, explaining how the specific question informs us about the big-picture question.
  • What is your research design? How will it answer the question? Usually it's best to address this by describing what the predictions are under different assumptions. I.e., "If the answer to my research question is X, I expect that this experiment will show the following results <show a graph here<, because of..... On the other hand, if the answer to my research question is Y, I expect that this experiment will show the following results...". To do this, you need clear linking hypotheses (i.e., explaining why Theory A makes us predict that pattern of results).
  • What were your research methods? i.e., you should tell me, in detail, the following things (the below list is not exhaustive; different kinds of research techniques may involve different things; just remember that you need to explain the methods clearly enough that someone else can exactly reproduce your study):
    • how many observations (participants, words, etc.) you used. (You should justify this number: why did you choose to have this many participants? In reality you will be limited by time and resource constraints this semester, but you can discuss how many participants you would use if you were able to do a full study)
    • what criteria the participants met (e.g. how old they were, what their language background was, etc.)
    • how many words/sentences/etc. you used (justify why)
    • what kinds of words/sentences/etc. you used (e.g. were there certain kinds of words you tried to avoid? or certain kinds of words that you always used? did you try to make sure that you had 10 of one kind of word and 10 of another kind of word? etc.)
    • What the experiment procedure was (i.e., exactly what the participants saw and did, with what timing)
    • how you analyzed the data. What specific part of data did you look at (e.g., how fast they respond to words; some specific kind of eye movement; whether they answer a question correctly or incorrectly; etc.)? Did you remove certain kinds of data (e.g., outliers? trials where people respond incorrectly?)
  • In cases where you made choices about how to do the experiment or how to analyze the data, you should justify those choices. e.g., why did you remove (or not remove) incorrect trials, outliers, etc.? Why did you show words at the speed you did? Why did you choose the corpus that you chose? and so on
  • What were your results? You should use a graph or table to illustrate the findings. Your graph should clearly show how the pattern varies (or does not vary) across participants. And your results must be accurate (i.e., if I analyze your data, I should be able to get the same results).
• You should include all relevant supplementary files. For example, if you created a DMDX script to collect reaction times, attach it to your report. If you measured phonetic data and put the results into Excel, attach the Excel file. If you wrote an R script to analyze your data, attach the script.
• Your proposal must be plagiarism-free. Most people don't understand what plagiarism actually is; therefore, to make sure you don't do anything that could be considered plagiarism, you should review the following document explaining plagiarism: plagiarism_and_attribution.pdf

If your submitted summary doesn't meet all of the above criteria, we will give you feedback and allow you one chance to revise. If your revised version does not meet all the criteria you will not be able to receive an A grade.

Recommendations

If you intend to attempt an A-level project, we advise you to complete the following modules:

• At least one of the following from Study Techniques...
  • "Reaction times" and "Using experiment control software"
  • "Phonetic analysis"
  • "Corpus analysis"
• All of the following from Data Analysis...
  • "Introduction to inferential statistics and t-tests"
  • "Study planning and evaluation"
  • "Programming"
  • "Data visualization"

by Stephen Politzer-Ahles. Last modified on 2021-05-02. CC-BY-4.0.